Module Descriptor School of Computer Science and Statistics
|Module Name||Linear Models II|
|Module Short Title||N/a|
|Semester Taught||Michaelmas term|
|Contact Hours||Lecture hours: 33Total hours: 33|
|Module Personnel||Lecturing staff: Prof Rozenn Dahyot|
When students have successfully completed this module they should be able to: Program, analyse and select the best model for explaining datasets. Interpret output of data analysis performed by a computer statistics package.
The aim of this module is to learn several mathematical techniques to analyse datasets for the purpose of explaining observed outcomes of experiments. Generalised Linear models (GLMs) are an extension to Standard Linear Regression. This extension is two fold. First, the distribution of the differences between the responses and their fitted values by the model is a member of the Exponential family (e.g. Normal, Poisson, Binomial etc.). Second, the relationship between the (expectation of the) responses and the exploratory variables is not anymore linear, and is chosen amongst several possible link (mathematical) functions. The course will focus on applying GLMs in several case studies using R.
Generalized Linear Models Exponential family AIC for model selection Deviance Generalised mixed linear models Multinomial distribution Survival Analysis With case studies analyzed with the R software
|Recommended Reading List|
|Module Prerequisites||Basic Statistics and Mathematics|
|Academic Year of Data||N/a|